Researchers at OpenAI decided that a system that scores well at understanding language could too easily be manipulated for malicious intent.

from Feed: All Latest http://bit.ly/2DCdMmk